Posts

Cadence: Building a Second Android App, and What I Changed

My first Android app was built to solve a problem only I had: monitoring my own infrastructure from my pocket. My second one was built to solve a problem someone else had, which turns out to be a very different kind of pressure. Cadence is an interval timer and reminder app, the kind of thing that nags you to stretch, take medication, or do a repeating task on a schedule, with proper quiet hours and a notification that won’t let itself be ignored. I built it for someone close to me who needed something more reliable than a generic timer app, and who, critically, would actually be using it every single day, not just admiring it from a developer’s chair.

Thursday, 9 July 2026 Read

The Config File That Wouldn't Stay Fixed

I’ve got a homelab service that’s been rock solid for months: a Cloudflare Tunnel connecting about thirty internal services to public subdomains, so I can reach everything from anywhere without opening a single port on my router. It’s the kind of infrastructure you stop thinking about, until it breaks twice in one week, for two completely different reasons, and both times the actual fault was hiding in a place I wasn’t looking.

Monday, 29 June 2026 Read

Fact-Checking 109 Ingredients With Two AI Models Before Shipping

Stagioni is a seasonal produce app I built with a partner in southern Italy: it covers around a hundred ingredients across vegetables, herbs, nuts, pulses, and fruit, each with a nutrition breakdown, a chef’s tip, a food science note, and storage guidance. It’s live on Google Play, in English and Italian, with a free tier and two paid tiers behind it. That last detail is the one that changed how I had to think about content quality. This isn’t a demo project where a wrong fact is embarrassing. It’s a published app making specific claims about nutrition and food science to paying users. If an entry says a compound does something it doesn’t, or cites a figure that’s been misremembered from somewhere, that’s not a typo. It’s misinformation with my name on it.

Sunday, 28 June 2026 Read

Giving My Homelab a Memory That Actually Understands What I Meant

My homelab has grown a knowledge base over time: sessions, lessons learned, backlog items, facts about how each piece of infrastructure is wired together. Hundreds of entries, all searchable, except “searchable” was doing a lot of quiet lying. The search was pure keyword matching: type the exact word that was in the original text, get a result; paraphrase it even slightly, get nothing. That’s a real problem when the whole point of the knowledge base is to stop me, or an AI assistant working alongside me, from repeating a mistake I’d already solved once. If the fix is recorded under one phrasing and I go looking with a different one, the knowledge base might as well not exist.

Tuesday, 16 June 2026 Read

Self-Healing Homelab MCP with AWS Route 53, CloudWatch and Lambda

There is a particular kind of frustration that comes from a tool dropping mid-session with no warning. My MCP server — the bridge between Claude and my homelab infrastructure — had developed a habit of going quiet at exactly the wrong moment. The fix I built turned out to be a solid piece of AWS architecture. The plot twist was that the problem itself was something else entirely. The Problem My MCP server runs on CT104 (ai-lab) inside a Proxmox LXC container on my home network, exposed via a Cloudflare tunnel. Claude connects to it over SSE (Server-Sent Events), which keeps a long-lived HTTP connection open so tool results can stream back in real time. When that connection dropped, the session was dead — no tools, no homelab access, nothing. The only recovery was a manual SSH in to restart the process.

Monday, 25 May 2026 Read

Building a Self-Hosted Voice-to-Text Pipeline with Whisper and AWS Bedrock

There is a small but persistent friction point in my workflow: I think faster than I type. When I am working through a problem or drafting something, the gap between thought and text slows everything down. I wanted to be able to speak, get clean written text back, and carry on without sending audio to a third-party API or paying per minute. This is how I built it. The Stack faster-whisper - a CTranslate2-based reimplementation of OpenAI’s Whisper. No PyTorch dependency, significantly smaller image footprint (~500MB vs ~2GB), and genuinely faster on CPU. FastAPI - the backend, handling audio uploads and browser recording. AWS Bedrock Nova Micro - an optional cleanup pass that fixes punctuation, capitalisation, and removes filler words from the raw transcript. Docker on Proxmox LXC - containerised on my existing Docker host alongside Pierre and Sablier. Cloudflare tunnel - HTTPS without a reverse proxy, accessible from desktop and mobile. Why Not the OpenAI Whisper API? The hosted API is fine for occasional use but charges per minute of audio. More importantly, audio stays local with a self-hosted solution. Everything runs on my own infrastructure - the only external call is the optional Bedrock cleanup step, which sends only the raw text transcript, not the audio itself.

Sunday, 26 April 2026 Read

How GitHub Became the Backbone of My Infrastructure

I started this week thinking of GitHub as a place to store code. I ended it realising it is infrastructure. Over 48 hours of building a self-hosted scraping engine, deploying a portfolio site to AWS, and recovering from a broken WireGuard tunnel, GitHub quietly became the thread that held everything together. Here is what I did not expect. Emergency Deployment My VPS went offline mid-project. No warning, no graceful shutdown, just gone. At that point, my portfolio site at anthonyapierre.com was running as a Hugo container on that same server.

Friday, 3 April 2026 Read

From Cloudflare Pages to AWS: Hosting My Portfolio on S3 + CloudFront

How I migrated my Hugo portfolio site from Cloudflare Pages to a production-grade AWS architecture using S3, CloudFront, ACM, IAM, and GitHub Actions — and what I learned along the way.

Wednesday, 1 April 2026 Read

When My VPS Went Down: Migrating to Cloudflare Workers in a Panic

I’ll be honest, I didn’t plan to migrate my site to Cloudflare Workers today. My hand was forced. I woke up to alerts telling me my VPS was down. My portfolio site, anthonyapierre.com, was completely unreachable. I’d been meaning to set up a proper failover for a while, but like most things that live on the “I’ll sort that eventually” list, it hadn’t happened yet. This is the story of how, on the very first day my site went fully live, an unexpected outage turned into something genuinely useful.

Monday, 30 March 2026 Read

Rebuilding the OSINT Lab: FastAPI, WebSockets, and 11 Tools Confirmed

When I first built the OSINT lab, the goal was simple: a dedicated, network-isolated environment for open source intelligence work running inside a Proxmox LXC container. Maltego and SpiderFoot were up and running, the container was air-gapped from the rest of the homelab, and it did the job. The problem was the dashboard. The original implementation used Node.js with Server-Sent Events to stream tool output to the browser, and it was fragile. SSE connections would drop unpredictably, tool output would get cut off mid-run, and the whole thing had a tendency to fall over under any kind of sustained load. It worked just well enough to be annoying rather than broken enough to force a fix — until I decided to fix it properly.

Saturday, 28 March 2026 Read

Building the HomelabDashboard Android App from Scratch

There is a certain point in homelab development where browser bookmarks and SSH sessions stop being a sensible way to manage things. When you are running 25+ LXC containers across a Proxmox host, a Romanian VPS, and a Docker stack, you want a single place to see everything — and you want it in your pocket. That is where HomelabDashboard came from. Not a third-party monitoring tool, not a pre-built solution — a native Android app built from scratch in Kotlin with Jetpack Compose, connected to a real Flask API running on my own infrastructure.

Tuesday, 24 March 2026 Read

Running an AI Monitoring Agent on AWS Bedrock for £0.02 a Month

When I started building an AI monitoring agent for the homelab, the obvious concern was cost. AWS Bedrock is powerful, but running AI analysis on a continuous basis against 25+ containers sounded like a recipe for a surprising monthly bill. The solution was tiered model routing — and it keeps the entire operation running for approximately £0.02 per month. The Tiering Logic AWS Bedrock’s Nova model family has three tiers with very different price points. Nova Micro is the cheapest available at around $0.035 per million input tokens. Nova Lite sits in the middle. Nova Pro is the most capable and the most expensive.

Wednesday, 18 March 2026 Read